The present invention generally relates to an image processing apparatus, and more particularly to an apparatus for providing a bilevel image signal corresponding to multilevel image data obtained from a scanned original image.
A scanner unit of a facsimile machine is an image processing apparatus for providing a bilevel image signal corresponding to multilevel image data obtained through raster scanning of an original image. The bilevel image signal is generated by subjecting the multilevel image data to a known binarization process by using a threshold value. Hereinafter, generating a bilevel image signal from multilevel image data is referred to as binarization. Several improved image processing devices of this type have been proposed for better picture quality of the bilevel image.
For example, Japanese Laid-Open Patent Application No. 56-109069 discloses a proposed image processing apparatus. In the conventional apparatus disclosed in the above mentioned publication, a bilevel image signal is generated from image data obtained from an original image by using the average of image data of the target pixel and its neighboring pixels as the threshold value of binarization.
However, when an original image containing an area of relatively uniform, intermediate density is read, the bilevel image corresponding to the uniform density area, produced by the conventional apparatus, may become unclear or dull. Also, the conventional apparatus may erroneously produce black pixels when a stained background portion of the original image is read and it is subjected to the binarization. In such cases, the resulting bilevel image contains some noise and the picture quality becomes poor. The conventional apparatus has some problems in that it does not detect whether or not a target pixel of the original image is in a uniform density area, and that it always uses the average of the image data as the threshold value of binarization even when the target pixel is in a uniform density area. In order to obtain better picture quality of the bilevel image, it is necessary to vary the threshold value of binarization depending on the features of the read areas of the original image.